Controle d'attitude d'un lanceur en phase atmospherique approche par applications gardiennes
Abstract
In a first phase, the modelling process underlines the presence of highly time varying parameters during the ascent, due to a fast mass variation along with propellant consumption. Linearizing the dynamical equations at six main flight instants yields linear time invariant models to be considered during control design. Each of them is to be stabilized by one control law, while respecting given specifications. The synthesis becomes even more complex when the bending modes are taken into account. Moreover, scheduling appears necessary to deal with the time variations. Indeed it is shown that no single gain setting is able to respect all the specifications along the trajectory. Furthermore, increasing complexity when modelling a whole launch vehicle pushes one to consider the model's errors and uncertainties. They represent a major issue in this study since it is asked to ensure the nominal performances in a robust fashion. Owing to their properties, guardian maps appear to be the most suitable tool to deal with such a problem of scheduling with robust performances. In light of this, the development of synthesis methods based on guardian maps is the main contribution of the project. It appears that actual state of the art in this field is focused on theoretical issues, whereas practical ones could be improved. Two approches are presented in the memoire. The first one is based on a graphical approach consisting in drawing the vanishing locus of guardian maps. A program using image analysis techniques is devised to check automatically which gain settings satisfy the constraints. The second one is based on an optimisation procedure involving guardian maps. Starting with the open loop system, the iterative process proposed ends up with a satisfactory gain setting for the closed-loop. These methods are tried and tested for the launch vehicle, with specifications from ASTRIUM-ST. Their practical application is motivated by the system complexity, the different kinds of constraints and the essential need for robustness. Many restrictions that finally bring about the interest and the efficiency of guardian maps for such a problem.
- Publication:
-
Masters Thesis
- Pub Date:
- 2012
- Bibcode:
- 2012MsT..........5D
- Keywords:
-
- Engineering, Aerospace;Engineering, Electronics and Electrical;Engineering, System Science